Advice on how to proceed with stress model

[Texas Piper]

Long time lurker, first time poster. I would greatly appreciate the input of the knowledgeable folks here with the following situation as I unfortunately do not have a mentor to turn to.

When doing stress analysis in the past, it was for newly installed systems with no tie-ins to existing lines so I am unsure of how to proceed here.

Project Background: We are relocating an existing pressure reducing valve station for 300# plant steam to allow the plant room for equipment additions. This also involves re-route of several plant steam lines, 300# and 100# as well as condensate lines. This piping ties into a main pipe rack.

Modeling and analyzing the entire existing system is outside our scope so how should I approach modeling the reaction at the existing tiepoints? I was thinking downstream of the tiepoints in the rack I would just include a straight run length of maybe 100’ with the associated supports, assuming it is restrained except for axial movement. The piping upstream of the 300# tiepoint is slightly more complex with a few directional and elevation changes. I was thinking of including the first few changes until it hits a ~100’ straight run out of the building again assuming it is restrained except for axial movement.

Attached is a plan view sketch to illustrate what I’m talking about. Please note that this in a very congested plant and the routing is done so to clear existing conditions.

Thank you for your input.

 

[KevinNZ]

Your model needs go out to anchors (or supports with line stops)

You will need to check if the new piping has changed the loads on the existing supports, so these need to be in the model. In some cases the loads from other side of anchors/supports needs to be added the modified side.

 

[Texas Piper]

Right, but realistically how much of the existing piping should be included in the stress analysis? The new piping will have new supports and it would seem there shouldn't be any significant load changes on the existing supports further down the rack. As posted above, I was going to include the first 100' of straight pipe in the rack with associated supports downstream of new piping. Then the first few directional changes until a straight run of the main 8"300# header upstream of new piping. Just looking for a sanity check, am I on the right path here?

 

[saplanti]

It depends on how you disturb the existing piping. As a pipe stresser you need to get the feeling that what will happen in case of new piping attachment, how they will interract.

In case the tie-in is close to the safety valves or a dynamic action both existing and new piping will go into motion. So you may have to extend your model until you find uneffected portion of the existing piping.

You perhaps need to talk to the process engineer for adequate location that will be motionless all the time. Otherwise you cannot get away from modeling the existing piping as explained above.

 

[saplanti]

My last paragraph first sentence is for the tie-in location. I have missed to put it there.

 

[LittleInch]

I don't get the bit about 100 ft of piping restrained except for axial movement??

Eh?

If you don't know that your existing line is moving or not then you can't model it properly.

IMO you need to find a point on the existing pipework where you are happy that the pipe isn't moving ( line stop or anchor) If you can't find one then you really don't have a point in space for the program to work out what is happening.

This happens a lot in plant / pipeline interface analysis where someone decides that as they don't know what is happening either side they either stick an enormous anchor there or have a "free end" which then shows large movements which are totally false as in reality it is connected to some large lumps of metal. The only way to model it properly is to do it as afar back as you need to before there is no movement. Everything else is fantasy. IMHO.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[1503-44]

If you post the stress iso maybe we can give better advice.

Reality used to affect the way we thought. Now we somehow believe that what we think affects reality.

 

[Texas Piper]

@LittleInch,

To clarify what I mean, the existing piping beyond our tiepoints is supported with pipe shoes with guides, so the piping cannot "move" or grow except in the axial direction. Downstream of our tiepoints, further down the rack, there are some directional changes and branch lines. I appreciate your advice on getting to a point where there is no movement. I guess ultimately what I was concerned about was where do you draw the line? To get the absolute BEST result you would need to model the entire existing system correct? And that was something that I just didn't see as being reasonable or even feasible with our scope of work and budget.

@7.0

I will be building the model soon, I will post once it is in reasonable shape to share.

 

[LittleInch]

Indeed, where do draw the line - not easy doing retrofit stuff, which is why sometimes you need to add in a line stop or anchor on the existing pipework where you don't think it is moving. There must be a line stop somewhere close? Or a big solid connection which acts line a line stop. The issue is that if you just state in the model that some point is a fixed location is that conservative or not? Same thing if you don't anchor it and just let the end of your 100 ft section float as a "free end" is that conservative or not? It's difficult to say without looking at the pipe routing and basically making a judgement call.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.